This invention relates generally to color digital cameras, and more particularly to the processing of pixel information generated by the sensor of a color digital camera.
Digital color cameras are used with computer or other digital processing systems. Such cameras include a sensor, optics, preprocessing electronics, and a cable or other communication link to transfer data to the digital processing system. Digital cameras are made by Connectix, Intel, and others.
In a conventional digital camera, the sensor is often a charge coupled device (CCD) that produces electrical image signals corresponding to an object producing or reflecting light onto the sensor. The electrical image signals are then processed and recorded on a storage medium such as a memory card or other computer readable medium.
A problem encountered with digital cameras is that the sensor can become saturated when the light intensity exceeds the intensity saturation limit of the sensor. When the light level is above the saturation level of the sensor, all further video information, other than the saturation information, is lost. In the past, this problem has been addressed by attempting to manipulate the amount of light striking the sensor or adjusting the sensor itself.
Some of the prior art has attempted to utilize a light intensity measuring circuit which processes data from the sensor and controls movement of an iris. The iris closes to restrict the amount of incident light striking the sensor and opens to allow more incident light to strike the sensor. However, the movement of the iris is slow compared to other means of intensity adjustment. Furthermore, due to its mechanical nature, the iris often closes more than necessary or doesn't close enough, at which point it must be repositioned. The result is slow correction time and fluctuations in brightness, ending in degradation of picture quality. The latter is especially if utilized with video cameras; where the fluctuating brightness is captured. Furthermore, a solitary bright light can cause the iris to close so much that gradation of the dark regions becomes compressed and deteriorated. The same occurs when direct strong light is incident, such as in the case of strong rear lighting.
In a system in which an electric charge corresponding to an amount of light received is accumulated on a photodiode and passed to an n-layer substrate of the sensor, other prior art has sought to draw away excess voltage that flows from the photodiode of the sensor when an excessive amount of light is received. For example, a p-layer is positioned between the photodiode and substrate of the sensor and grounded. This applies a reverse bias voltage to the substrate and p-layer so that a depletion layer is formed between the photodiode and the substrate. Surplus electric charges that overflow from the photodiode, due to an excessive amount of received light, are absorbed in the depletion layer. The voltage of the substrate is then adjusted to accommodate bright and dark scenes. A great disadvantage is that signals from the photodiodes are disrupted in that they must now pass through a layer specifically designed to absorb such signals. Another disadvantage of this is that the voltage of the substrate must be preset based on estimated light conditions, especially disadvantageous for capturing moving video. Another disadvantage is that the voltage of the entire substrate must be changed, not just for the portion receiving the excessively intense light. This results in poor picture quality in that bright areas of light are compensated for but dim spots are not.
Another problem encountered with digital cameras is that picture quality drops as the light intensity falls below a certain level. Furthermore, even images in a dark portion of a scene can be hard to observe if a direct strong light is incident in the scene. The prior art has attempted to correct such problems by monitoring signals from the sensor that have been separated into red, green and blue component signals and have also been gamma corrected. Then, a dark area proportion in a whole picked up image is detected. Next, a portion of the processed original signal is modified to stretch the dark signal region to improve the gradation of the dark area. The processed original signal and the modified signal are combined to output a resultant gradation improved signal. The disadvantage of this prior art method is that separation and gamma correction of the original signal are performed before the dark area is detected. This reduces the accuracy of the detection of the dark area. Further, the dark area is estimated from the processed original signal as a whole, not on a pixel by pixel basis, further reducing the accuracy of the detection of the dark area and making correction of the dark area more difficult.